Bienvenido: Ingresar

Por favor indica el password de tu cuenta en el wiki remoto abajo.
/!\ Deberías confiar en ambos wikis porque los administradoresrespectivos podrían ver tu contraseña.

Quitar mensaje
location: JavierAndresRedolfi / sartb

Fisher Vectors for PolSAR Image Classification

Abstract

In this letter we study the application of the Fisher Vector (FV) to the problem of pixel-wise supervised classification of PolSAR images. This is a challenging problem since information in those images is encoded as complex-valued covariance matrices. We observe that the real part of these matrices preserve the positive semidefiniteness property of their complex counterpart. Based on this observation, we derive a FV from a mixture of real Wishart pdfs and integrate it with a Potts-like energy model in order to capture spatial dependencies between neighboring regions. Experimental results on two challenging datasets show the effectiveness of the approach.

Paper

Javier Redolfi, Jorge Sánchez, and Ana Georgina Flesia
"Fisher Vectors for PolSAR Image Classification"
IEEE Geoscience and Remote Sensing Letters, 99, Septiembre 2017. (link)

Code

sartb.tar.gz

How to install.

Dataset

For evaluation, we consider a subset of the images that were available trough PolSARpro by the European Space Agency (ESA). This subset consists of two fully polarimetric images over the San Francisco Bay (SFB) area, USA, and over an agricultural region in the Flevoland (FL) province in The Netherlands.

dataset

San Francisco Bay

Pseudocolor representation of the polarimetric data (left) and the ground truth labels (right).

sfb.png

Flevoland

Pseudocolor representation of the polarimetric data (left) and the ground truth labels (right).

flev.png

Classification Results

Mean accuracy on SFB and FL for the CWC, RWM and the FV-based approaches


Dataset

CWC

RWM ($K$=32, $n$=3)

Ours ($K$=16, $n$=5)

SFB

0.8848 (0.0602)

0.9081 (0.0238)

0.9707 (0.0079)

FL

0.8762 (0.0223)

0.8608 (0.0312)

0.9144 (0.0191)


Classification results on the SFB (left) and FL (right) images


sfb_segmentation.png fl_segmentation.png

Comparison with the best performing method of [1] in FL


Method

BareSoil

Beet

Forest

Grasses

Lucerne

Peas

Potatoes

Rapeseed

Beans

Water

Wheat

Mean

Wang etal. [1]

0.9878

0.9371

0.9496

0.8489

0.9231

0.9555

0.8896

0.9486

0.9653

0.9642

0.8864

0.9324

Ours ($K=16, n=5$)

1.0

0.9926

0.9770

0.9920

0.9261

0.9993

0.9959

0.9983

0.9757

0.8472

0.9971

0.9728


Comparison with [2] in FL


Method

Accuracy

W-DSN [2]

0.9268 (-)

Ours ($K$=16, $n$=5)

0.9688 (0.0062)


References

[1] H. Wang, Z. Zhou, J. Turnbull, Q. Song, and F. Qi, “Pol-SAR classification based on generalized polar decomposition of mueller matrix,” IEEE Geosc. Remote Sens. Lett., vol. 13, no. 4, pp. 565–569, 2016.

[2] L. Jiao and F. Liu, “Wishart deep stacking network for fast polsar image classification,” IEEE Trans. Image Process., vol. 25, no. 7, pp. 3273–3286, 2016.